SPSS On-Line Training Workshop

Home
Table of Contents
Overview of Analysis
Data Editor
Output
Program Editor
Chart Editor
Analysis
Projects

Statistical Procedure

In this Tutorial:

runs

Nonparametric tests are runs, 1-sample Kolmogorov-Smirnov, Independent samples and related samples.

In this on-line workshop, you will find many movie clips. Each movie clip will demonstrate some specific usage of SPSS.

Runs - This is used when you want to test for randomness

	Cut point - this is the value that you choose (median, mode, mean, other) to divide the variable into two separate groups.
	Options - If you want descriptives for your variable, you must choose this.

1-Sample K-S - This is used to test if the variables have a specified distribution

Test Distribution - We must choose from the four tests listed:

	Normal - compare our distribution to the Normal distribution
	Poisson - compare our distribution to the Poisson distribution
	uniform - compare our distribution to the Uniform distribution
	exponential - compare our distribution to the Exponential distribution

Options - If we want descriptives for our variable, we must choose this.

2 Independent Samples - this test is similar to the independent t-test, but it is done when the distribution is not normal.

	Test Type - Mann-Whitney is the default and is the most common. If you want to test for both location and shapes, you would choose the Kolmogorov-Smirnov or the Wald-Wolfowitz.
	Options - If you want descriptives for your variable, you must choose this.

K Independent Samples - this test is similar to ANOVA, but is done when the distribution is not normal

	Test type - Kruskal-Wallis is the default, and is one of the most popular tests.
	Options - If we want descriptives for our variable, we must choose this.

2 Related Samples - this is similar to the paired or related t-test, but is done when the distribution is not normal.

	Test type - Wilcoxon and Sign are for continuous data. (Wilcoxon is more powerful.) If you have binary data, the McNemar test is the test to choose.
	Options - If you want descriptives for your variable, you must choose this.

K Related Samples - this is similar to the repeated measures ANOVA, but is done when the distribution is not normal.

	Test Type - Friedman's is a popular test. Kendall's is a normalization of the Friedman test. Cochran's is used for binary data.
	Statistics - If you want descriptives for your variable, you must choose this.